trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.942339373970346,
  "global_step": 3000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.82,
      "learning_rate": 4.5881383855024714e-05,
      "loss": 0.5588,
      "step": 500
    },
    {
      "epoch": 1.65,
      "learning_rate": 4.1762767710049426e-05,
      "loss": 0.4,
      "step": 1000
    },
    {
      "epoch": 2.47,
      "learning_rate": 3.764415156507414e-05,
      "loss": 0.3519,
      "step": 1500
    },
    {
      "epoch": 3.29,
      "learning_rate": 3.352553542009885e-05,
      "loss": 0.3179,
      "step": 2000
    },
    {
      "epoch": 4.12,
      "learning_rate": 2.9406919275123562e-05,
      "loss": 0.2926,
      "step": 2500
    },
    {
      "epoch": 4.94,
      "learning_rate": 2.528830313014827e-05,
      "loss": 0.2696,
      "step": 3000
    }
  ],
  "max_steps": 6070,
  "num_train_epochs": 10,
  "total_flos": 5476078662709248.0,
  "trial_name": null,
  "trial_params": null
}